The RISE system: conquering without separating
Author:
- 17 December 2002
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 704-707
Abstract
Current rule induction systems (e.g. CN2) typically rely on a "separate and conquer" strategy, learning each rule only from still-uncovered examples. This results in a dwindling number of examples being available for learning successive rules, adversely affecting the system's accuracy. An alternative is to learn all rules simultaneously, using the entire training set for each. This approach is implemented in the RISE 1.0 system. Empirical comparison of RISE with CN2 suggests that "conquering without separating" performs similarly to its counterpart in simple domains, but achieves increasingly substantial gains in accuracy as the domain difficulty grows.Keywords
This publication has 5 references indexed in Scilit:
- A further comparison of splitting rules for decision-tree inductionMachine Learning, 1992
- Rule induction with CN2: Some recent improvements1991
- The CN2 induction algorithmMachine Learning, 1989
- Induction of decision treesMachine Learning, 1986
- A theory and methodology of inductive learningArtificial Intelligence, 1983